Venture capital investments in artificial intelligence (AI), one of the biggest technological trends these days, are booming. The year 2016 saw almost ten times more funds invested in the space than 2012 did. Virtual and augmented reality (AR/VR), sometimes associated with the creative side of AI, also became a hot topic: More than $1.8 billion was invested in AR/VR, compared with just $86 million in 2012. Applying AI to solve tasks normally considered creative has historical precedent. For example, procedural generation has been used to draw textures, produce 3D models, and automatically generate large amounts of content in video games since the 1980s.

Chronic obstructive pulmonary disease (COPD) is a lung disease where early detection benefits the survival rate. COPD can be quantified by classifying patches of computed tomography images, and combining patch labels into an overall diagnosis for the image. As labeled patches are often not available, image labels are propagated to the patches, incorrectly labeling healthy patches in COPD patients as being affected by the disease. We approach quantification of COPD from lung images as a multiple instance learning (MIL) problem, which is more suitable for such weakly labeled data. We investigate various MIL assumptions in the context of COPD and show that although a concept region with COPD-related disease patterns is present, considering the whole distribution of lung tissue patches improves the performance. The best method is based on averaging instances and obtains an AUC of 0.742, which is higher than the previously reported best of 0.713 on the same dataset. Using the full training set further increases performance to 0.776, which is significantly higher (DeLong test) than previous results.

We present mlrMBO, a flexible and comprehensive R toolbox for model-based optimization (MBO), also known as Bayesian optimization, which addresses the problem of expensive black-box optimization by approximating the given objective function through a surrogate regression model. It is designed for both single- and multi-objective optimization with mixed continuous, categorical and conditional parameters. Additional features include multi-point batch proposal, parallelization, visualization, logging and error-handling. mlrMBO is implemented in a modular fashion, such that single components can be easily replaced or adapted by the user for specific use cases, e.g., any regression learner from the mlr toolbox for machine learning can be used, and infill criteria and infill optimizers are easily exchangeable. We empirically demonstrate that mlrMBO provides state-of-the-art performance by comparing it on different benchmark scenarios against a wide range of other optimizers, including DiceOptim, rBayesianOptimization, SPOT, SMAC, Spearmint, and Hyperopt.

In this paper, we aim at recovering an undirected weighted graph of $N$ vertices from the knowledge of a perturbed version of the eigenspaces of its adjacency matrix $W$. For instance, this situation arises for stationary signals on graphs or for Markov chains observed at random times. Our approach is based on minimizing a cost function given by the Frobenius norm of the commutator $\mathsf{A} \mathsf{B}-\mathsf{B} \mathsf{A}$ between symmetric matrices $\mathsf{A}$ and $\mathsf{B}$. In the Erd\H{o}s-R\'enyi model with no self-loops, we show that identifiability (i.e., the ability to reconstruct $W$ from the knowledge of its eigenspaces) follows a sharp phase transition on the expected number of edges with threshold function $N\log N/2$. Given an estimation of the eigenspaces based on a $n$-sample, we provide support selection procedures from theoretical and practical point of views. In particular, when deleting an edge from the active support, our study unveils that our test statistic is the order of $\mathcal O(1/n)$ when we overestimate the true support and lower bounded by a positive constant when the estimated support is smaller than the true support. This feature leads to a powerful practical support estimation procedure. Simulated and real life numerical experiments assert our new methodology.

Supervised learning has been very successful for automatic segmentation of images from a single scanner. However, several papers report deteriorated performances when using classifiers trained on images from one scanner to segment images from other scanners. We propose a transfer learning classifier that adapts to differences between training and test images. This method uses a weighted ensemble of classifiers trained on individual images. The weight of each classifier is determined by the similarity between its training image and the test image. We examine three unsupervised similarity measures, which can be used in scenarios where no labeled data from a newly introduced scanner or scanning protocol is available. The measures are based on a divergence, a bag distance, and on estimating the labels with a clustering procedure. These measures are asymmetric. We study whether the asymmetry can improve classification. Out of the three similarity measures, the bag similarity measure is the most robust across different studies and achieves excellent results on four brain tissue segmentation datasets and three white matter lesion segmentation datasets, acquired at different centers and with different scanners and scanning protocols. We show that the asymmetry can indeed be informative, and that computing the similarity from the test image to the training images is more appropriate than the opposite direction.

Acquiring your first language is an incredible feat and not easily duplicated. Learning to communicate using nothing but a few pictureless books, a corpus, would likely be impossible even for humans. Nevertheless, this is the dominating approach in most natural language processing today. As an alternative, we propose the use of situated interactions between agents as a driving force for communication, and the framework of Deep Recurrent Q-Networks for evolving a shared language grounded in the provided environment. We task the agents with interactive image search in the form of the game Guess Who?. The images from the game provide a non trivial environment for the agents to discuss and a natural grounding for the concepts they decide to encode in their communication. Our experiments show that the agents learn not only to encode physical concepts in their words, i.e. grounding, but also that the agents learn to hold a multi-step dialogue remembering the state of the dialogue from step to step.

Chatbots are exploding in the finance industry, and it's no surprise why. With the ability to automate operations, reach more customers, and provide a more friction-free banking experience, chatbots are streamlining and optimizing many banks' digital services. Here's how the world's biggest banks are using chatbots to boost their business. As one of the largest U.S. banks, Bank of America (BofA) is leading the charge for artificially intelligent chatbots in financial services. Last year, the company announced Erica, a voice- and text-enabled chatbot for BofA customers.

Since my childhood, I have been working towards my goal of building human-level AI. But I had two main areas of interest – game programming and artificial intelligence. I spent my days and nights developing various real-time algorithms for computer graphics, aiming for a career in game development. This period deepened my understanding of programming and computer architectures. I learned how to invent new algorithms, to optimize the code on a machine level, to use math and physics to simulate the world, to develop game engines and to do art, game design, and sound design.

For many years, computers could play mathematical games like chess and the sheer computing horsepower was the key to their success. Last year, for the first time ever, a computer was able to play the complex board game of Go and win! "AlphaGo", the artificially intelligent player developed by Google, won the game by defying millennia of human instinct and intellect. What is most interesting about this is that it was not preprogrammed to play Go: rather, it learned using a general-purpose algorithm which allowed it to interpret the game's patterns and came up with entirely new ways of approaching the game that originated in China more than two thousand years ago. Clearly, machine learning and the larger world of artificial intelligence (AI) are no longer the stuff of science fiction. They're here–and many businesses are already taking advantage.

Researchers have overcome one of the major stumbling blocks in artificial intelligence with a program that can learn one task after another using skills it acquires on the way. Developed by Google's AI company, DeepMind, the program has taken on a range of different tasks and performed almost as well as a human. Crucially, and uniquely, the AI does not forget how it solved past problems, and uses the knowledge to tackle new ones. The AI is not capable of the general intelligence that humans draw on when they are faced with new challenges; its use of past lessons is more limited. But the work shows a way around a problem that had to be solved if researchers are ever to build so-called artificial general intelligence (AGI) machines that match human intelligence.